BACKGROUND: Dairy products are high in saturated fat and are traditionally a risk factor for vascular diseases. The fatty acids 15:0 and 17:0 of plasma lipids are biomarkers of milk fat intake. The aim of the present study was to evaluate the risk of a first-ever stroke in relation to the plasma milk fat biomarkers. METHODS: A prospective case-control study was nested within two population based health surveys in Northern Sweden. Among 129 stroke cases and 257 matched controls, plasma samples for fatty acid analyses were available in 108 cases and 216 control subjects. Proportions of 15:0 and 17:0 of plasma lipids, weight, height, blood lipids, blood pressures, and lifestyle data were employed in conditional logistic regression modelling. RESULTS: The proportions of fatty acids 17:0 and 15:0+17:0 of total plasma phospholipids were significantly higher in female controls than cases, but not in men. 17:0 and 15:0+17:0 were significantly and inversely related to stroke in the whole study sample as well as in women. The standardised odds ratio (95% CI) in women to have a stroke was 0.41 (0.24-0.69) for 17:0 in plasma phospholipids. Adjustment for traditional cardiovascular risk factors, physical activity and diet had marginal effects on the odds ratios. A similar, but non-significant, trend was seen in men. CONCLUSION: It is hypothesised that dairy or milk fat intake may be inversely related to the risk of a first event of stroke. The intriguing results of this study should be interpreted with caution. Follow up studies with greater power, and where intakes are monitored both by dietary recordings and fatty acid markers are needed.